Alteration of blood-brain barrier integrity by retroviral infection

Abstract

The blood-brain barrier (BBB), which forms the interface between the blood and the cerebral parenchyma, has been shown to be disrupted during retroviral-associated neuromyelopathies. Human T Lymphotropic Virus (HTLV-1) Associated Myelopathy/Tropical Spastic Paraparesis (HAM/TSP) is a slowly progressive neurodegenerative disease associated with BBB breakdown. The BBB is composed of three cell types: endothelial cells, pericytes and astrocytes. Although astrocytes have been shown to be infected by HTLV-1, until now, little was known about the susceptibility of BBB endothelial cells to HTLV-1 infection and the impact of such an infection on BBB function. We first demonstrated that human cerebral endothelial cells express the receptors for HTLV-1 (GLUT-1, Neuropilin-1 and heparan sulfate proteoglycans), both in vitro, in a human cerebral endothelial cell line, and ex vivo, on spinal cord autopsy sections from HAM/TSP and non-infected control cases. In situ hybridization revealed HTLV-1 transcripts associated with the vasculature in HAM/TSP. We were able to confirm that the endothelial cells could be productively infected in vitro by HTLV-1 and that blocking of either HSPGs, Neuropilin 1 or Glut1 inhibits this process. The expression of the tight-junction proteins within the HTLV-1 infected endothelial cells was altered. These cells were no longer able to form a functional barrier, since BBB permeability and lymphocyte passage through the monolayer of endothelial cells were increased. This work constitutes the first report of susceptibility of human cerebral endothelial cells to HTLV-1 infection, with implications for HTLV-1 passage through the BBB and subsequent deregulation of the central nervous system homeostasis. We propose that the susceptibility of cerebral endothelial cells to retroviral infection and subsequent BBB dysfunction is an important aspect of HAM/TSP pathogenesis and should be considered in the design of future therapeutics strategies.

Figure 1. Expression of HTLV-1 receptors, Glut-1 and NP-1, in the thoracic spinal cord from an uninfected individual.

(A,B) Glut-1 immunostaining by immunoperoxydase technique (iPO) (DAB substrate). (C) Dual staining for Glut-1 (iPO, DAB substrate, brown color) and factor VIII (Alcaline phosphatase (AP), FastBlue substrate, blue color). (D,E) NP-1 staining by iPO technique (DAB substrate). Thoracic spinal cord specimens were cut in a cryostat, fixed in methanol and processed for iPO as described in Materials and Methods. (F) Dual staining for NP-1 (iPO, DAB substrate, brown color) and factor VIII (Alcaline phosphatase, FastBlue substrate, blue color). Frozen tissue samples from the thoracic spinal cord were cut on a cryostat at 10 µm, fixed in methanol and processed for immunohistochemistry as described in Materials and Methods. Magnification: (A,D) 5×, (B–E) 50×, (C–F) 100×.

Figure 2. Detection of Glut-1 and HTLV-1 transcripts in the thoracic spinal cord from HAM/TSP patient.

(A,B,C) Glut-1 staining by iPO (DAB substrate, brown) on thoracic spinal cord sections from a HAM/TSP patient. Nuclei were counterstained with Harri's haematoxilin solution (blue). (A,B) Cryostat section; (C) paraffin section. Magnification: (A) 30×; (B,C) 75×. (D,E) Detection of HTLV-1 transcripts (tax) in cryosections of the spinal cord by in situ hybridization. Arrow heads indicate positive cells and vascular structures. Astrocytes were detected by iPO (DAB substrate, brown) against GFAP. Magnification: 220×.

Figure 3. Expression of viral receptors at the cell surface of hCMEC/D3 cells and susceptibility to infection by HTLV-Env pseudotyped particles.

(A) FACS analysis of Glut-1 expression on hCMEC/D3 cells. Signal for permeabilized cells appear as bold-lined curve; unpermeabilized as thin-lined curve; control cells (without Glut-1 antibody) appear as a grey-filled curve. (B) FACS analysis of Neuropilin-1 expression on hCMEC/D3 cells. Signal for unpermeabilized as bold-lined curve; control cells (without NP-1 antibody) appear as a grey-filled curve. (C) FACS analysis of Heparan sulfate-proteoglycans expression on hCMEC/D3 cells. Signal for unpermeabilized as bold-lined curve; control cells (without HSPGs antibody) appear as a grey-filled curve. (D) Infection assay on hCMEC/D3 cells using Lac-Z-carrying MLV particles pseudotyped with the HTLV (HTLV) or Amphotropic (MLV) envelope proteins. β-gal production was measured using chemiluminescence after subtracting the background of uninfected cells on total protein normalization. Infections were performed without serum, or in the presence of serum from a HAM/TSP patient (HAM/TSP) (1/10) or from a HTLV-uninfected individual (control) (1/10). Results represented as means±SD.

Figure 4. Syncytia formation between hCMEC/D3 cells and HTLV-1 infected MT-2 lymphocytes at 24 h post-contact.

(A) Role of the viral proteins in the formation of the syncytia. Evaluation of the number and size (nb of nuclei/syncytium) of the syncytia obtained by coculture between hCMEC/D3 and MT-2 cells in the presence of serum from a HAM/TSP patient (1/7) (noted HAM/TSP) or from an uninfected individual (control, noted HTLV negative control or in the absence of any serum (w/o serum). (B) Role of the viral receptors in the formation of the syncytia. Evaluation of the number and size (nb of nuclei/syncytium) of the syncytia obtained by coculture between hCMEC/D3 and MT-2 cells in the presence of dextran sulfate in blue (that prevents the HTLV-1 Env/HSPG interaction), or in the presence of VEGF165 in green (that prevents the HTLV-1 Env/NRP-1 interaction), or in the presence of a polyclonal antibody against Glut-1 in red (that prevents the HTLV-1 Env/GLUT1 interaction). Results are representative of 3 independent experiments. (C) Detection of viral p24 protein (green) by immunofluorescence in a syncytium. Nuclei were stained with DAPI (blue). (D) Demonstration of endothelial origin of the syncytia by prior labeling of hCMEC/D3 cells with a red vital fluorescent marker (Cell-tracker, red). Nuclei were stained with DAPI (blue). Magnification (B–C): 350×.

Figure 5. Productive infection of hCMEC/D3 cells by HTLV-1.

(A) Kinetics of p19 viral protein secretion in the supernatant of hCMEC/D3 and irradiated HTLV-1 MT-2 lymphocyte cocultures. Cells were cultivated or not in presence of 25 µM AZT. Results are mean and standard deviation from triplicate experiments. (B) Kinetics of detection of infected hCMEC/D3 cells by irradiated MT-2 cells. Infection was assessed by FACS analysis of p24 viral protein at days 12, 14, 16 and 22 post-coculture. (C) Production of infectious viral particles by HTLV-1-infected endothelial cells using a reporter cell-line (293T-LTR-GFP). The supernatant of hCMEC/D3 infected cells was collected and ultracentrifuged and the resuspended pellet was applied on the reporter cell-line 293T-LTR-GFP. The expression of GFP was assessed 6 days later after cell fixation.

Figure 6. Alteration of BBB functions in the infected hCMEC/D3 cells.

(A) Altered permeability of a monolayer of infected hCMEC/D3 cells. The endothelial cells were cocultured with irradiated MT-2 or C81-66 lymphocytes for 15 days. Then hCMEC/D3 cells were seeded on Transwell filters and permeability to FITC-dextran 70 kDa was assessed after differentiation of the monolayer. (B) Effect of endothelial cells infection on CEM lymphocyte migration. Infected endothelial cells were seeded onto filters. The migration was estimated at 24 hours of culture by fluorescence assay after labeling of lymphocytes (from the CEM, Jurkat, MT-2 of C81-66 cell-lines) with a fluorescent marker. The migration rate is expressed as ratio (%) of fluorescence intensity in the lower compartment versus total fluorescence. (C) Analysis of the expression level of ZO-1, Occludin, and viral p24 of a monolayer of hCMEC/D3 cells cocultured with irradiated MT-2 or C81-66 for 15 days. β-tubulin was used for normalization. (D) Inhibiting MLC phosphorylation has no effect on the permeability for FITC dextran 70 kDa across a monolayer of infected endothelial cells. hCMEC/D3 cells were seeded on filters. After differentiation, cultures were either left untreated of treated for 48 hours with ML-1, a specific inhibitor for MLCK activity. Permeability was then estimated as described in Materials and Methods.